[dplthuy]

Hi all,

I want to model the breakup of a liquid column under the influence of a step change in pressure & velocity (shock wave), using the compressibleInterIsoFoam (VOF) solver. The goal of the case is to determine a minimum grid cell size to properly resolve the breakup behavior, by comparison to literature data (https://doi.org/10.2514/6.2017-1892).

The simulations are two-dimensional. I am not yet taking into account the presence of the liquid (i.e. alpha.liquid = 0 in the whole domain), just considering the propagation of the pressure wave through the domain. The domain is initialised with p=p_rgh=8.9e4 Pa (gravity is neglected), U = (0 0 0).
On the left side of the domain, I create an area with higher pressure, p=p_rgh=1.09e5 Pa, U = (80 0 0), as can be seen in attached picture, for t=0. Left domain boundary is an inlet with these values specified with fixedValue BC. Top and bottom boundaries are zeroGradient for all parameters.
After the pressure wave passes through the domain, I would like it to exit at the right outlet boundary. However, as shown in the picture, the pressure wave reflects back into the domain.

Initially, I used zeroGradient conditions for p, p_rgh, and U at the outlet. Since this resulted in wave reflection, I have also tried the advective and waveTransmissive for p, p_rgh and U. I have also applied pressureInletOutletVelocity or zeroGradient for U in combination with advective or waveTransmissive for p and p_rgh. None of these different conditions made any difference in the wave reflection.

I know that density-based solvers (such as rhoCentralFoam), are better suited for pressure wave modeling as compared to pressure-based solvers such as I am using. Indeed, when using rhoCentralFoam in combination with zeroGradient outlet condition for p and U for my case-setup, there is no wave reflection at the outlet. However, since I want to use this case to determine a proper VOF resolution for droplet breakup, I do not have the option of switching to a density-based solver.

Is it possible at all to avoid these pressure wave reflections when using a pressure-based solver? If so, which combination of boundary conditions is recommendable?

I hope I have provided enough information on my case, if not, please let me know what is missing. Thanks in advance!

[JulioPieri]

Try using something different for U. waveTransmissive can be paired with zeroGrad in U, for instance. I'm not sure you can use waveTransmissive for p_rgh: I'm facing this very problem right now.